Breast cancer represents a heterogeneous disease process with markedly different treatment outcomes despite similar clinical staging. Even with the most aggressive and modern therapy as much as 15 percent of patients develop local-regional recurrences (LRR) which increase the risk of distant disease and death. Individual risk factors such as age, tumor size and grade are clinically limited in ability to predict outcome; therefore, the identification of a molecular profile to more accurately assess risk remains an important goal. This profile can help determine LRR as part of the evaluation for local therapeutic options, particularly when adjuvant treatment recommendations supported by traditional risk stratification parameters result in a narrow risk: benefit ratio. Previous evaluation of a genetic profile for LRR has been hindered by smaller patient numbers with heterogeneous clinical characteristics and treatment regimens. Additional study is warranted to develop a profile that would not only predict LRR but be applicable to patients regardless of clinical factors such as tumor size or age. The need for such a tool is greatest, in patients currently characterized as being at intermediate risk, i.e. stage T1-T3 breast cancer with 0-3 positive nodes. Patients treated with mastectomy without neoadjuvant therapy and without postoperative radiation therapy would provide the most uniform subset of patients since their pathologic staging is not altered by neoadjuvant therapy. We hypothesize that the molecular profile of patients with a LRR differ than the molecular profile of patients who do not have a recurrence. A retrospective case-control study will be performed, matching two control patients without LRR for each patient with a post-mastectomy LRR. Our aims are 1) To determine molecular aberrations (DNA, RNA, IHC) associated with an elevated risk of LRR in otherwise low to intermediate risk post-mastectomy patients; 2) To determine whether PI3K pathway activation by PIK3CA mutation and/or PTEN loss is associated with LRR; 3) To determine if tumor proliferation, as determined by Ki-67, is associated with LRR. Our long-term goal is to create a novel molecular profile that would predict LRR in this cohort and then further validate this signature in 1) a separate TBCRC (Translational Breast Cancer Research Consortium) cohort in a follow-up study, 2) in paraffin blocks stored from appropriate cooperative group trials, 3) with future TBCRC patient trials as a correlative and translational component. Upon validation in an independent cohort, we will pursue a prospective study to use our LRR signature for determining the need for adjuvant radiation therapy. A molecular profile may provide more information to guide an individualized therapeutic approach to breast cancer therapy. Then clinical management would be based on a women's individual prognosis using the tumor's molecular profile to define the best therapy.